Kidney mitochondrial DNA contributes to systemic IL-6 release in sepsis-associated acute kidney injury

Abstract

Mitochondrial dysfunction is a major mechanism of acute kidney injury (AKI), and increased circulating interleukin 6 (IL-6) is associated with systemic inflammation and death due to sepsis. We tested whether kidney mitochondrial DNA (mtDNA) contributes to IL-6 release in sepsis-associated AKI via Toll-like receptor 9 (TLR9). In a murine model of sepsis via cecal ligation and puncture (CLP), we used next-generation sequencing of plasma mtDNA to inform the design of optimal target sequences for quantification by droplet digital PCR, and to identify single-nucleotide polymorphisms (SNPs) to infer tissue origin. We found significantly higher concentrations of plasma mtDNA after CLP versus shams and that plasma mtDNA SNPs matched kidney SNPs more than other organs. Kidney mtDNA contributed directly to IL-6 and mtDNA release from dendritic cells in vitro and kidney mitochondria solution led to higher IL-6 concentrations in vivo. IL-6 release was mitigated by a TLR9 inhibitor. Finally, plasma mtDNA was significantly higher in septic patients with AKI compared with those without AKI and correlated significantly with plasma IL-6. We conclude that AKI contributes to increased circulating IL-6 in sepsis via mtDNA release. Targeting kidney mitochondria and mtDNA release are potential translational avenues to decrease mortality from sepsis-associated AKI.

Keywords: Inflammation; Innate immunity; Mitochondria; Nephrology.

Figure 1. Systemic cytokines are increased prior to plasma creatinine in cecal ligation and puncture models of sepsis-associated AKI.

Plasma creatinine (A) was significantly elevated 24 hours after cecal ligation and puncture (CLP) compared with sham operation, but no significant differences were found at the 4-hour time point. Conversely, TNF-α (B), KC/GRO (C), IL-1β (D), IL-10 (E) increased and remained increased compared with sham at 24 hours and 4 hours. IL-2 (F), IL-5 (G), IL-6 (H), and IL-12p70 (I) increased at 4 hours, but returned to sham levels at 24 hours. Plasma levels of IFN-γ (J) decreased at 24 hours in CLP compared with sham. Significant differences between groups were determined by 1-way ANOVA with Fisher’s LSD post hoc analysis. A, n = 6–7/group; B–J, n = 4–6/group.

Figure 2. Circulating cell-free DNA fragments are increased after cecal ligation and puncture compared with sham operation and circulating cell-free mtDNA is highly fragmented.

DNA was isolated from mouse plasma 4 hours after cecal ligation and puncture (CLP) versus sham operation. Next-generation sequencing was used to determine the size distribution of the circulating cell-free mtDNA fragments. Minimum Hamming distance for the entire mouse mitochondrial genome compared to the nuclear genome was determined by blastn in the ungapped mode. (A) The concentrations of total cell-free DNA fragments between 100 and 150 bp in length were determined by automated electrophoresis and found to be increased significantly in CLP mice versus shams. (B) Size distribution analysis of circulating cell-free mtDNA fragments showed a high degree of fragmentation after CLP, with roughly 40% of all fragments being less than 150 bp and roughly 15% being less than 100 bp in length. The overall size distribution of circulating cell-free mtDNA fragments did not differ significantly between CLP mice versus shams. (C) Several regions of the mouse mitochondrial genome, such as within the cytochrome c oxidase (COX) subunits, have very low minimum Hamming distances compared with the nuclear genome, indicating a high degree of similarity to a region of nuclear DNA. Significant differences between groups were determined by unpaired, 2-tailed t tests comparing the group means ± SEM. A, n = 4/group; B, n = 5/group. CDF, cumulative distribution function.

Figure 3. Circulating cell-free mtDNA is increased in mice after cecal ligation and puncture compared with sham operation.

Plasma concentrations of mtDNA were determined by droplet digital PCR targeting sequences within the cytochrome b (CytB) and NADH dehydrogenase 1 (ND1) genes. These genes were chosen based on their dissimilarity from any region of the nuclear genome based on minimum Hamming distance, and small sequences were targeted to account for fragmentation. Plasma concentrations of CytB (A), ND1a (B), and ND1b (C) were all increased significantly in cecal ligation and puncture (CLP) mice compared with shams at 4 hours. CytB and ND1b remained significantly increased at 24 hours compared with shams, and CytB was also increased significantly at 4 hours versus 24 hours after CLP. (D) The smallest target sequence, CytB, was increased significantly compared with ND1a and ND1b at 4 hours after CLP. (E) Simple linear regression between plasma concentrations of CytB and ND1a and ND1b, respectively, showed a positive and significant correlation for both comparisons. Significant differences between groups were determined by 1-way ANOVA with Tukey’s post hoc analysis. n = 6/group.

Figure 4. The kidney has increased mtDNA single-nucleotide polymorphisms.

When compared with the heart, liver, lung, and buffy coat, the kidney had an absolute increase in single-nucleotide polymorphisms (SNPs) (P = 0.0068, P = 0.0216, P = 0.0049, P = 0.001 respectively). Significant differences between groups were determined by 1-way ANOVA with Fisher’s LSD post hoc analysis. n = 5/group.

Figure 5. Kidney metabolomic changes are increased compared with other organs after sepsis by cecal ligation and puncture.

Next-generation metabolomics was used to evaluate metabolic changes in the kidney after cecal ligation and puncture (CLP) compared to those in the lung, liver, and heart. (A) PLS-DA, a supervised multivariate analysis, showing a significant difference in the metabolomic profile in kidney tissue of mice after CLP versus sham operation. (B) Volcano plot analysis shows the number of significantly increased or decreased metabolites in kidney tissue of mice after CLP versus sham operation, with 20% of the 483 total metabolites analyzed showing significant differences. Fold change greater than 1.2 and FDR-adjusted P value of less than 0.05 were considered significant. (C) Top 25 metabolites contributing to metabolomic differences between CLP versus sham groups as ranked by VIP scores showed majority of the most significantly altered metabolites were phospholipids, i.e., phosphatidylcholine [PC(32:2)], or medium- and long-chain acylcarnitines (i.e., adipoylcarnitine). (D) Comparison of the sum of the fractional impact of all significantly altered metabolites shows the greatest sum in the kidney followed by the lung, heart, and liver, respectively. n = 4 for sham, n = 5 for CLP.

Figure 6. CpG in vitro in DCs and kidney mtDAMP in vivo contributes to IL-6 release.

(A) DCs isolated from mouse bone marrow were treated with CpG (1 μM) as a positive control or kidney mtDNA (10 μg/mL) and incubated for 4 hours prior to quantification of IL-6 in the culture supernatant. Additional cells were additionally treated with TLR9 inhibitor. Cells treated with CpG and mtDNA had a significant increase in IL-6 in supernatant compared to controls. However, IL-6 concentrations in supernatants from cells treated with mtDNA and a TLR9 inhibitor decreased significantly from mtDNA treatment alone. (B) Mice were treated with mtDAMPs followed by measurement of plasma IL-6 at 4 hours with and without TLR9 inhibition. Plasma IL-6 concentrations were increased significantly in mice at 4 hours following subcutaneous injection of mtDAMPs vs normal saline (control). Plasma IL-6 concentrations decreased significantly after mtDNA plus TLR9 inhibition compared mtDNA alone. (C) mtDNA levels were measurable via CYTB. Significant differences between groups were determined by 1-way ANOVA with Fisher’s LSD (A and B) or Tukey’s (C) post hoc analysis. Experiment in A was completed 3 separate times. A, n = 5 wells/group; B, n = 4/group.

Figure 7. Circulating cell-free mtDNA concentrations are increased in individuals with sepsis-associated AKI compared with those with sepsis without AKI and correlate with plasma IL-6.

mtDNA was quantified by droplet digital PCR targeting a 69-bp target sequence within the NADH dehydrogenase 1 (ND1) gene in plasma from a cohort of individuals with sepsis with and without acute kidney injury (AKI). (A) Plasma ND1 concentrations were increased significantly in individuals who developed AKI within 48 hours of ICU admission (AKI) compared with those who did not (no AKI). (B) IL-6 concentrations in septic individuals with AKI were significantly increased compared with those without AKI. Significant differences between groups were determined by unpaired, 2-tailed t tests comparing the group means ± SEM. (C) Simple linear regression of plasma ND1 with IL-6 concentrations in individuals with sepsis showed a positive and significant correlation. n = 31 patients.